Abstract: A method and apparatus for matching a first sequence of patterns representative of a first signal with a second sequence of patterns representative of a second signal using a dynamic programming matching technique is described. The second signal patterns which are at the end of a dynamic programming path for a current first signal pattern are listed in an active list 201. The dynamic programming paths are propagated by processing the second signal patterns on the active list, and a new active list 205 is generated for the succeeding input pattern. In order to propagate each path, the system determines how many second signal patterns lie within an overlap region in which a comparison has to be made, and processes each path in dependence upon the determined amount of overlap.

Abstract: A method and apparatus which is used to precondition a speech signal such that the signal has relatively low power at predetermined points which form the boundaries of DFT blocks in a coder. The method and apparatus is particularly effective when the filter bank operates on a linear-prediction residual. The requirement of having low energy at the block boundary is well approximated by a requirement of having a pitch pulse near the center of the block. The method and apparatus makes it possible to make the difference between the original speech signal and the pre-processed speech signal inaudible or nearly inaudible. An AE coder which follows the pre-processor, therefore, reconstructs a quantized version of the pre-processed speech. The present invention differs from earlier pre-processors in its operation, in the properties of the modified speech signal, and in the fact that it is compatible with a sinusoidal or waveform-interpolation type of speech coder.

Abstract: The word processor of the present invention comprises: a voice inputting device for inputting spoken word and converting the spoken word into voice data; a voice storage device for storing the voice data; a speech recognition device for recognizing a word in the voice data output from the voice inputting device or the voice data stored by the voice storage device; a display for displaying a result obtained by the voice recognition device; an instruction inputting device for inputting an instruction to select a portion in the result; and a correction device for correcting the portion in the result according to the instruction from the instruction inputting device.

Abstract: A method and apparatus for signal classification using a multilayer temporal relaxation network involves receiving an input signal feature vector, classifying a first signal feature, and classifying a second signal feature using contextual information. The multilayer temporal relaxation network applies a relaxation process that updates an activation value of a node in a first layer and updates an activation value of a node in a second layer. The multilayer network then generates a signal classification according to an activation value of a node in the multilayer network.

Abstract: An adaptive speech recognition system is provided including an input for receiving a signal derived from a spoken utterance indicative of a certain vocabulary item, a speech recognition dictionary, a speech recognition unit and an adaptation module. The speech recognition dictionary has a plurality of vocabulary items each being associated to a respective dictionary transcription group. The speech recognition unit is in an operative relationship with the speech recognition dictionary and selects a certain vocabulary item from the speech recognition dictionary as being a likely match to the signal received at the input. The results of the speech recognition process are provided to the adaptation module. The adaptation module includes a transcriptions bank having a plurality of orthographic groups, each including a plurality of transcriptions associated with a common vocabulary item.

Abstract: The system performs unsupervised speech model adaptation using the recognizer to generate the N-best solutions for an input utterance. Each of these N-best solutions is tested by a reliable information extraction process. Reliable information is extracted by a weighting technique based on likelihood scores generated by the recognizer, or by a non-linear thresholding function. The system may be used in a single pass implementation or iteratively in a multi-pass implementation.

Abstract: A speech synthesis apparatus synthesize a speech signal by filtering a speech source signal through a synthesis filter. A speech source signal codebook stores a plurality of speech source signals as a code vector. A unit dictionary memory stores a plurality of synthesis units corresponding to phonemic symbols, each synthesis unit comprising an index of the code vector in the speech source codebook and a shift number for the code vector to decode the speech source signal. A unit selection section selects a synthesis unit corresponding to phonemic symbols to be synthesized from the unit dictionary memory. A synthesis unit decoder selects the code vector corresponding to the index in the synthesis unit from the speech source signal codebook, and shifts the code vector according to the shift number in the synthesis unit.

Abstract: Speech is encoded into a frame of bits. A speech signal is digitized into a sequence of digital speech samples that are then divided into a sequence of subframes. A set of model parameters is estimated for each subframe. The model parameters include a set of voicing metrics that represent voicing information for the subframe. Two or more subframes from the sequence of subframes are designated as corresponding to a frame. The voicing metrics from the subframes within the frame are jointly quantized. The joint quantization includes forming predicted voicing information from the quantized voicing information from the previous frame, computing the residual parameters as the difference between the voicing information and the predicted voicing information, combining the residual parameters from both of the subframes within the frame, and quantizing the combined residual parameters into a set of encoded voicing information bits which are included in the frame of bits.

Abstract: A camera is targeted using voice recognition analysis. Audio information is received by a talker identification (TID) module from a microphone. The TID module automatically performs a voice recognition analysis on the audio information to uniquely identify which of a plurality of talkers is talking. The camera is automatically controlled to target a camera preset location corresponding to the talker identified to be talking.

Abstract: An apparatus capable of, and a method of, playing audio, the apparatus comprising communicating, processing, and playing means for, and the method comprising the steps of: communicating a user's information preferences to an information provider; receiving, from the information provider, informational items that are responsive to the user's information references; interleaving and sequencing, for the user, a playing of the received informational items with a playing of a plurality of musical items included in an audio library of the user; and playing, for the user and responsive to the interleaving and sequencing, the received informational items within a playing of the plurality of musical items; and wherein the playing comprises a voice synthesizing of an at least one of informational item; wherein the playing is responsive to a schedule preferences of the user; wherein a verified apparent listening of a playing of an informational item is associated with a credit; and/or wherein a user's rec

Abstract: A system for adaptively generating a composite noisy speech model to process speech in, e.g., a nonstationary environment comprises a speech recognizer, a re-estimation circuit, a combiner circuit, a classifier circuit, and a discrimination circuit. In particular, the speech recognizer generates frames of current input utterances based on received speech data and determines which of the generated frames are aligned with noisy states to produce a current noise model. The re-estimation circuit re-estimates the produced current noise model by interpolating the number of frames in the current noise model with parameters from a previous noise model. The combiner circuit combines the parameters of the current noise model with model parameters of a corresponding current clean speech model to generate model parameters of a composite noisy speech model. The classifier circuit determines a discrimination function by generating a weighted PMC HMM model.

Abstract: An improved, more intelligent network interface unit capable of processing signals received from a high speed transmission media and outputting the received signals without substantial amplification or reduction. The network interface unit comprises a regenerator circuit which detects T1 signals from said high-speed transmission media and outputs a digital stream to a serial data processor for processing via a micro-controller. A wave shaper circuit receives the output from the serial data processor and, in conjunction with the micro-controller and a level detector circuit, regenerates signals having substantially the same wave shape and amplitude as the original T1 signals received from said transmission media.

Abstract: A method and apparatus for generating frame voicing decisions for an incoming speech signal having periods of active voice and non-active voice for a speech encoder in a speech communications system. A predetermined set of parameters is extracted from the incoming speech signal, including a pitch gain and a pitch lag. A frame voicing decision is made for each frame of the incoming speech signal according to values calculated from the extracted parameters. The predetermined set of parameters further includes a frame full band energy, and a set of spectral parameters called Line Spectral Frequencies (LSF).

Abstract: A method (100) of compressing a digital signal that is parametrically modeled and encoded includes the steps of storing (102) the digital signal in a memory in a plurality of frames having a plurality of parameters in each frame of the plurality of frames, wherein the digital signal was encoded at a higher rate and converting the digital signal to a lower rate by selecting (106) from each frame of the plurality of frames a subset of the plurality of parameters and discarding (108) the subset of the plurality of parameters within each frame of the plurality of frames.

Abstract: An apparatus and a method for imaging the mouth area laterally to produce reliable measurements of mouth and lip shapes for use in assisting the speech recognition task. A video camera is arranged with a headset and a microphone to capture a lateral profile image of a speaker. The lateral profile image is then used to compute features such as lip separation, lip shape and intrusion depth parameters. The parameters are used in real time, during speech recognition process to characterize and discriminate spoken phonemes to produce a high degree of accuracy in automatic speech recognition processing, especially in a noisy environment.

Abstract: A system and method for indexing an audio stream for subsequent information retrieval and for skimming, gisting, and summarizing the audio stream includes using special audio prefiltering such that only relevant speech segments that are generated by a speech recognition engine are indexed. Specific indexing features are disclosed that improve the precision and recall of an information retrieval system used after indexing for word spotting. The invention includes rendering the audio stream into intervals, with each interval including one or more segments. For each segment of an interval it is determined whether the segment exhibits one or more predetermined audio features such as a particular range of zero crossing rates, a particular range of energy, and a particular range of spectral energy concentration. The audio features are heuristically determined to represent respective audio events including silence, music, speech, and speech on music.

Abstract: An audio coding system encodes and decodes audio signals as a plurality of independent layers of coded audio data. A basic representation of the original audio signal may be reconstructed from decoding of a single layer of coded audio data. However, a more complete representation of the original audio signal is reconstructed by decoding additional layers of coded audio data. The coding system finds application with decoding systems of varying processing power, and in transmission systems having communication channels that are characterized by intermittent transmission errors and/or variable capacity. At an encoding system, an audio signal is broken into a plurality of frequency bands which are filtered, down sampled and independently coded. A decoding system inverts the coding process applied at the encoding system for whatever number of layers that is determined will be decoded.

Abstract: A method and system of providing network based transcription of free-form speech signals. A speech signal is recorded as a digital audio file in a storage medium and is then streamed over a data network to a client terminal for transcription. As the speech signal arrives, it is buffered in memory at the client terminal while a streaming player application plays the signal to a transcriptionist. The transcriptionist then conveniently listens to and transcribes the speech signal as it is being played. The invention advantageously avoids the need to physically transfer and download the full digital audio to a transcriptionist computer or to transport physical storage media, such as tapes or CD-ROM from the place of recording to the place where the recorded voice signals will be transcribed.

Abstract: A method and apparatus are provided for automatically acquiring grammar fragments for recognizing and understanding fluently spoken language. Grammar fragments representing a set of syntactically and semantically similar phrases may be generated using three probability distributions: of succeeding words, of preceding words, and of associated call-types. The similarity between phrases may be measured by applying Kullback-Leibler distance to these three probability distributions. Phrases being close in all three distances may be clustered into a grammar fragment.

Abstract: A digital isolation apparatus for coupling between a telecommunication network and a digital signal processing device. A CODEC is arranged to couple to the telecommunication network, convert a first set of analog signals from the telecommunication network to a first set of digital signals, and convert a second set of digital signals to a second set of analog signals to the telecommunication network. A digital isolation device, arranged to connect between the CODEC and the digital signal processing device, pass the first set of digital signals from the CODEC to the digital signal processing device, and pass the second set of digital signals from the digital signal processing device to the CODEC. A power supply assembly supplies power to the CODEC via the digital isolation device.